[petsc-dev] Fwd: SACUDA - Cannot compile

Barry Smith bsmith at mcs.anl.gov
Wed Oct 27 21:59:27 CDT 2010


On Oct 27, 2010, at 9:43 PM, Victor Minden wrote:

> Satish: That patch seemed to work (everything compiled anyway), thanks.
> 
> Barry: undef before the include and define again afterwards worked also.
> 
> Is one preferable?

   Yes, do the undef before and the def again after.  Satish's approach is totally disgusting.

   Barry

> 
> Victor
> ---
> Victor L. Minden
> 
> Tufts University
> School of Engineering
> Class of 2012
> 
> 
> On Wed, Oct 27, 2010 at 9:24 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:
> 
> On Oct 27, 2010, at 5:31 PM, Jed Brown wrote:
> 
> > On Wed, Oct 27, 2010 at 18:24, Satish Balay <balay at mcs.anl.gov> wrote:
> > Ok - perhaps we can simulate current behavior with:
> >
> > typedef char* VecType;
> > typedef const char* ConstVecType;
> >
> > and use VecType/ConstVecType as needed in code?
> 
>  Yuck, no way. Over my dead body.
> 
> >
> > This or just use the typedef in the CUDA interfaces (after #undef VecType).  Dirty, but would fix the current collision.
> 
>   Why not do this? Just undef VecType before the cuda/thrust include and def it again afterwards.
> 
>   Barry
> 
> >
> > Jed
> 
> 




More information about the petsc-dev mailing list